Relative risk of glucose elevation during antipsychotic exposure in a Veterans Administration population.

Concern is mounting that atypical antipsychotics cause disturbance in glucose regulation ranging from reversible hyperglycemia to diabetic ketoacidosis and death. It is difficult, however, to know what the level of risk of hyperglycemia might be for an individual patient on a particular medication of this class. We conducted a retrospective nonrandomized cohort analysis of glucose measurements in 18,764 patients receiving outpatient prescriptions for olanzapine, risperidone, or typical antipsychotics from 1 October 1998 to 30 June 2003 at six Veterans Affairs Medical Centers in the southeast United States. In patients without a random plasma glucose measurement > or =160 mg/dl before medication exposure (n=1394), treatment with index medications was associated with an incidence of new diabetes-level hyperglycemia of 78.7 cases per 1,000 individuals exposed per year. Olanzapine exposure was associated with a greater rate of developing at least one glucose measurement > or =200 mg/dl than risperidone (odds ratio=2.14, P=0.003). Olanzapine exposure was also associated with a greater rate of development of at least one fasting glucose measurement > or =126 mg/dl than risperidone. Typical antipsychotics were associated with risk intermediate between the two atypicals. These data indicate that patients with no previously observed glucose elevations develop diabetes-level hyperglycemia during antipsychotic treatment, particularly in patients receiving olanzapine.